Download Factor-Deficient Mice Colony-Stimulating Granulocyte

Protection from Collagen-Induced Arthritis in
Granulocyte-Macrophage Colony-Stimulating
Factor-Deficient Mice
This information is current as
of June 18, 2017.
Ian K. Campbell, Melissa J. Rich, Robert J. Bischof, Ashley
R. Dunn, Dianne Grail and John A. Hamilton
J Immunol 1998; 161:3639-3644; ;
http://www.jimmunol.org/content/161/7/3639
Subscription
Permissions
Email Alerts
This article cites 35 articles, 23 of which you can access for free at:
http://www.jimmunol.org/content/161/7/3639.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 1998 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
References
Protection from Collagen-Induced Arthritis in
Granulocyte-Macrophage Colony-Stimulating
Factor-Deficient Mice1
Ian K. Campbell,2* Melissa J. Rich,* Robert J. Bischof,* Ashley R. Dunn,† Dianne Grail,† and
John A. Hamilton*
G
ranulocyte-macrophage CSF (GM-CSF)3 is a glycoprotein traditionally viewed as a growth and differentiation
factor necessary for the development of hemopoietic
progenitor cells into granulocytes, macrophages, and dendritic
cells (1, 2). However, in view of its diverse actions on mature
hemopoietic cells, it has been suggested that GM-CSF may also be
a proinflammatory cytokine (3–5). Notably, GM-CSF has been
reported to have the following actions: induction of class II MHC
expression and urokinase-type plasminogen activator production
by monocyte-macrophages (3, 5), enhancement of granulocyte and
monocyte cellular adherence (4, 6), augmentation of macrophage
APC function (7), priming of monocytes for cytokine production
(8, 9), stimulation of phagocytosis and superoxide production by
neutrophils (10, 11), and neutrophil chemotaxis (12).
Rheumatoid arthritis (RA) is a chronic inflammatory autoimmune disease for which there is evidence that GM-CSF may be
involved. GM-CSF has been found at elevated levels in RA lesions
(13) and is produced in vitro by resident joint cells (chondrocytes
*Department of Medicine, Inflammation Research Center, University of Melbourne,
and †The Ludwig Institute for Cancer Research, Royal Melbourne Hospital, Parkville,
Victoria, Australia
Received for publication February 12, 1998. Accepted for publication May 26, 1998.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by a Program Grant from the National Health and Medical
Research Council of Australia and in part by grants from the Arthritis Foundation of
Australia and AMGEN Boulder Inc. (Boulder, CO).
2
Address correspondence and reprint requests to Dr. Ian K. Campbell, Autoimmunity
and Transplantation Division, The Walter and Eliza Hall Institute, P.O. Royal Melbourne Hospital, Victoria 3050, Australia. E-mail address: [email protected]
3
Abbreviations used in this paper: GM-CSF, granulocyte-macrophage CSF; RA,
rheumatoid arthritis; CIA, collagen-induced arthritis; CII, collagen type II; DTH,
delayed-type hypersensitivity.
Copyright © 1998 by The American Association of Immunologists
and synovial fibroblasts) following their stimulation with inflammatory cytokines, such as IL-1 and TNF-a (14, 15). The latter
observations led to the CSF network hypothesis (16) that sought to
explain the chronicity of rheumatoid joint disease in terms of a
positive feedback between joint cell CSF secretion and monokine
production. GM-CSF has also been implicated in the adherence of
neutrophils to cartilage and its subsequent degradation by these
cells (17). Finally, GM-CSF has been reported to cause a flare-up
of existing RA when administered (for correction of neutropenia)
to patients with Felty’s syndrome or following chemotherapy
treatment (18, 19).
Collagen-induced arthritis (CIA) in the mouse (20) is an autoimmune model of RA that is dependent upon both humoral and
cellular immune responses to type II collagen (CII) (21); it is considered to be restricted to mouse strains bearing the H-2q or H-2r
haplotypes and is generally performed in DBA/1 mice (22). We
recently reported that GM-CSF, when injected i.p. into DBA/1
mice suboptimally primed to develop CIA, exacerbated the disease
symptoms (23), suggesting a proinflammatory role for GM-CSF in
this model. Although highlighting the importance that elevated
circulating levels of GM-CSF could have on the course of disease,
this study did not indicate whether endogenous GM-CSF was a
necessary component of the CIA response.
Therefore, in the present study we have examined the role of endogenous GM-CSF in the CIA model using GM-CSF-deficient mice
(24). For this purpose we have taken the unique approach of establishing the model in mice of a non-H-2q/non-H-2r background,
thereby eliminating the need for backcrossing the GM-CSF-deficient
mice onto the DBA/1 (H-2q) strain. We report that GM-CSF-deficient
mice are relatively resistant to the induction of CIA compared with
their littermate wild-type control mice. This study implicates GMCSF as a key proinflammatory cytokine pivotal to the development of
CIA in mice and adds further support to the idea of the involvement
of GM-CSF in inflammatory joint diseases (e.g., RA).
0022-1767/98/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
The involvement of granulocyte-macrophage CSF (GM-CSF) in collagen-induced arthritis (CIA) was examined using GM-CSFdeficient mice. Although CIA is generally considered to be restricted to mice of the H-2q or H-2r haplotypes, we examined the role
of GM-CSF in the CIA model using GM-CSF-deficient (2/2) and wild-type (1/1) mice on a C57BL/6 (H-2b) background. Mice
were immunized by intradermal injection at the base of the tail with chick type II collagen followed by a repeat injection 21 days
later. We found, based on both clinical and histologic assessments, that wild-type mice on this background developed severe CIA,
while the GM-CSF-deficient mice had virtually no disease. Mice that were heterozygous for the GM-CSF gene (1/2) collectively
displayed an intermediate response between those of the GM-CSF1/1 and GM-CSF2/2 groups, suggesting a gene dosage effect.
GM-CSF1/1 and GM-CSF1/2 mice exhibited CIA responses ranging from mild (single digits) to severe swelling of all four paws,
while in the few GM-CSF2/2 mice that developed CIA the disease was confined to single digits. Despite the putative role of
GM-CSF in dendritic cell development, GM-CSF-deficient mice exhibited both humoral and cellular (delayed-type hypersensitivity) responses to type II collagen; however, the cellular response was significantly reduced in the GM-CSF-deficient mice
compared with the wild-type controls. These findings suggest that GM-CSF is required for CIA development in mice and support
the idea that GM-CSF is a key cytokine in inflammatory joint disease. The Journal of Immunology, 1998, 161: 3639 –3644.
3640
PROTECTION FROM CIA IN GM-CSF NULL MICE
Materials and Methods
Mice
Mice heterozygous for a disrupted GM-CSF gene (GM-CSF1/2) were provided by the Ludwig Institute for Cancer Research (Royal Melbourne Hospital, Melbourne, Australia) and bred at the Department of Medicine animal house. The derivation was previously reported (24). Briefly, chimeric
mice were generated by microinjection of 129/OLA-derived ES cells (H2b) with a disrupted GM-CSF gene into C57BL/6 (H-2b) host blastocysts.
Germline transmitters of the mutated GM-CSF allele were crossed with
C57BL/6 mice for 11 generations, giving GM-CSF1/2 mice that were
interbred to yield the GM-CSF2/2, GM-CSF1/2, and GM-CSF1/1 mice
used for the experiments. In some experiments (see text) mice from the first
C57BL/6 cross were interbred and used; these were confirmed as homozygous H-2b by FACS analysis of spleen cells using H-2b-specific mAb (I. K.
Campbell and P. M. Hogarth, unpublished observations). GM-CSF genotype status was determined by PCR analysis of tail DNA as previously
described (24). Animals were fed standard rodent chow and water ad libitum and were housed with same sex littermates in sawdust-lined cages.
Mice of both sexes were consigned to experiments at 8 to 15 wk of age.
Collagen-induced arthritis
Clinical and histologic assessment of arthritis
Animals were assessed for redness and swelling of limbs, and a clinical
score was allocated for each mouse two to three times per week for up to
60 days as previously described (23). The maximum score per mouse was
12. At termination, the rear paws of the mice were removed, fixed, decalcified, and paraffin embedded as previously described (23). Frontal sections
(5 mm) were stained with hematoxylin and eosin and evaluated without
knowledge of the treatment groups, based on the histologic assessment of
Williams et al. (25).
ELISA for detection of Abs to CII
ELISA assays were performed for the detection of Abs to CII by coating
96-well flat-bottom plates (Immunoplate Maxisorp, Nunc, Copenhagen,
Denmark) with 50 ml of CII (2 mg/ml in PBS) overnight at 4°C. The wells
were then blocked by 1-h incubation at 37°C with 200 ml of PBS containing 1% (w/v) BSA. Next, 50 ml of serial fourfold dilutions (beginning at
1/1000 dilution) of mouse sera in PBS supplemented with 0.05% (v/v)
Tween-20 were applied and incubated for 2 h at 37°C. Horseradish peroxidase-conjugated goat anti-mouse whole IgG (Sigma) or isotype-specific
(IgG1, IgG2a, IgG2b, and IgG3; Southern Biotechnology, Birmingham,
AL) antisera (50 ml) were next applied for 2 h at 37°C followed by ophenylenediamine dihydrochloride substrate (Sigma) in phosphate-citrate
buffer (50 ml), and color development was monitored after a standard period by measurement in a microplate reader (model 450, Bio-Rad, Richmond, CA) at 450 nm. Three washes in PBS/0.05% (v/v) Tween-20 were
applied between all steps. Standard curves were constructed as follows: for
total IgG, protein G-Sepharose affinity-purified mouse anti-CII IgG fraction (serial 1/4 dilutions beginning at 7 mg/ml); for others, mouse anti-CII
sera (serial 1/4 dilutions beginning at 1/1000). In each case the anti-CII sera
were derived from a pool obtained from CII-hyperimmunized DBA/1 mice.
Arbitrary units were assigned to the standards, such that 1 U/ml gave an
OD of 0.5 with the different antisera.
Delayed-type hypersensitivity (DTH) reaction
Mice were immunized, as before, by intradermal injection into the base of
the tail with chick CII in CFA. Ten days later the mice were anesthetized
and injected s.c. into the right hind footpad with 20 ml of a solution containing 2 mg/ml CII in PBS; the left footpad received the same volume of
vehicle. Immediately before injection and 24 and 48 h thereafter, the thicknesses of the left and right footpads were measured using spring callipers
(Mitutoyo, Tokyo, Japan) accurate to 0.01 mm. The Ag-specific DTH response was determined as the increase in right footpad thickness minus the
increase in left footpad thickness at the given time points. Injection of CII
into the paws of naive mice produced negligible swelling. Following the
48 h measurement, the mice were sacrificed, and the hind footpads were
FIGURE 1. CIA development in GM-CSF-deficient mice and littermate controls. The incidence of arthritis (shown as cumulative percentage;
a) and the mean clinical scores (6SEM; b) of GM-CSF2/2 (n 5 15),
GM-CSF1/2 (n 5 28), and GM-CSF1/1 (n 5 13) mice are shown with
time following primary immunization with CII. The final incidence of arthritis was significantly lower in the GM-CSF2/2 mice than in the GMCSF1/1 and GM-CSF 1/2 mice (p , 0.005 for each, by x2 test). For
statistical analysis of clinical scores see Table I.
removed and processed for histologic analysis as described above, except
that decalcified specimens were halved in the sagittal plane, and paraffinembedded sections were cut in this plane from the center of the footpad
outwards. Sections were stained with hematoxylin and eosin to confirm
cellular infiltration into the dermis and s.c. regions.
Statistics
For clinical scores the Mann-Whitney two-sample rank test was used to
determine the level of significance between means of groups. For data
pertaining to Ab levels in serum samples and the DTH reaction, Student’s
t test for the difference between two means was employed, while the paired
t test was used for comparison of mouse weights over the course of the
experiments. The incidence of arthritis between different groups and the
proportion of joints in different histologic categories were assessed by the
x2 test. For each test p , 0.05 was considered statistically significant.
Results
Reduced incidence and severity of CIA in GM-CSF-deficient
mice
To evaluate the requirement for endogenous GM-CSF for the development of CIA, GM-CSF-deficient mice (GM-CSF2/2) and
their littermate controls (GM-CSF1/1 and GM-CSF1/2), each on
the C57BL/6 (H-2b) background, were primed to develop CIA by
intradermal immunization with chick CII in CFA followed by a
repeat of the primary injection 21 days later. This immunization
schedule successfully elicits CIA in the wild types of these and
certain other non-H-2q/non-H-2r mouse strains with an incidence
only slightly lower than that in DBA/1 mice (our manuscript in
preparation).
Figure 1 shows results from the pooled data of two experiments
performed using 8- to 14-wk-old mice from the 11th C57BL/6
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
An emulsion was formed by dissolving 2 mg/ml chick CII (Sigma, St.
Louis, MO) overnight at 4°C in 10 mM acetic acid and combining it with
an equal volume of CFA containing 5 mg/ml heat-killed Mycobacterium
tuberculosis (H37 Ra, Difco, Detroit, MI). Mice were injected intradermally at several sites into the base of the tail with a total of 100 ml of
emulsion containing 100 mg of CII; this was repeated as a boost 21 days
later. Mice immunized without the CII component did not develop arthritis
during the period of investigation in this study.
The Journal of Immunology
3641
Table I. Response of GM-CSF 1/1, 1/2, and 2/2 mice to CIA
Table II. Histopathological assessment of joints from GM-CSF1/1 and
GM-CSF2/2 micea
GM-CSF Genotype
Parameter
1/1
1/2
GM-CSF Genotypeb
2/2
Clinical assessmenta
Average clinical
3.4 6 1.0 2.1 6 0.5 0.1 6 0.1e
score (d22-60)b
Severity median (range)c 9 (1–12)
6 (2–12)
2 (1–2)
Anti-CII IgG (mg/ml)d
1.60 6 0.32 2.04 6 0.35 1.71 6 0.29
a
Summary of data in Figure 1.
Values show mean 6 SEM of the clinical scores of individual mice averaged
over days 22 to 60.
c
Values show arthritis severity as the median (and range) of maximum clinical
scores obtained for the responsive mice of each genotype.
d
Values are mean 6 SEM concentrations of anti-CII total IgG in the mouse sera
at day 60.
e
p , 0.005 and p , 0.001, compared to GM-CSF1/1 and GM-CSF1/2 groups,
respectively.
b
1/1
2/2
Total joints
Normal
Mild
Moderate
Severe
57
7 (12)
3 (5)
6 (11)
41 (72)
57
52 (91)c
1 (2)
2 (4)
2 (4)c
a
Joints of clinically positive mice in the GM-CSF1/1 (n 5 4) and GM-CSF2/2
groups (n 5 2, plus two other mice) were examined histologically.
b
Data show the number (% in parentheses) of joints distributed within each histological severity grade for the two genotypes.
c
p , 0.001, compared to GM-CSF1/1 group.
the responsive GM-CSF1/1 mice had a broad range from 1 to 12
(median 9); the means (6SEM) of the clinical scores averaged
over days 23– 60 differed significantly for GM-CSF1/1 and GMCSF2/2 mice (2.18 6 0.79 and 0.02 6 0.02, respectively;
p , 0.01).
Histologic assessment
To confirm the clinical assessments, at sacrifice the clinically positive hind paws of the two responding GM-CSF2/2 mice as well as
the hind limbs of two other GM-CSF2/2 mice and those of four
clinically positive GM-CSF1/1 mice from the experiments detailed in Figure 1 were removed and processed for paraffin embedding and sectioning. Histologic grading of hematoxylin- and
eosin-stained sections was performed for each joint based on the
procedure of Williams et al. (25), and the results are summarized
in Table II. The GM-CSF2/2 mice exhibited a significantly reduced proportion of joints in the severe histopathologic category
compared with that in the GM-CSF1/1 mice (4 vs 72%, respectively; p , 0.001) and were more frequently normal in histologic
appearance (91% compared with 12% for GM-CSF1/1 mice; p ,
0.001). The GM-CSF2/2 mouse joints were typically normal in
appearance (Fig. 2a), while inflammatory cell infiltrate of granulocytes and mononuclear cells, subsynovial inflammation and hyperplasia, as well as cartilage and bone degradation resulting in
loss of joint architecture were common features of the GMCSF1/1 mouse joints (Fig. 2, b and c).
In the majority of GM-CSF2/2 mouse joints examined (52 of
57) there was a complete absence of pathology (Fig. 2a). However, the interphalangeal and metatarsal-phalangeal joints of the
digits showing clinical signs of arthritis in some instances (2 of
57 joints) exhibited histopathology comparable to that observed
in the most severely affected GM-CSF1/1 mice. Inflammatory
cell infiltrate, synovial hyperplasia, pannus formation, and associated cartilage and bone degradation were evident with loss
of joint architecture.
Humoral response to CII of GM-CSF-deficient mice
Since CIA development is dependent on both B and T cell responses (21), the almost complete absence of CIA in the GMCSF2/2 mice could be due either to their inability to develop Abs
to CII or to a weakened cellular response to CII. To address the
first possibility, sera were collected from all mice at sacrifice and
examined by ELISA for Abs to CII. The total IgG Abs to CII were
comparable for the different mouse genotypes (Table I). Since the
IgG2a and IgG2b isotypes, rather than the total IgG, are considered
to be of prime importance in the Ab response to CII (27), sera from
GM-CSF2/2 and GM-CSF1/1 mice were also examined for possible differences in the IgG isotype responses to CII. This further
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
backcross (99.95% C57BL/6); in these experiments the entire litters of GM-CSF1/2 cross-matings were tested. Immunized mice
were examined regularly over a 60-day period for signs of redness
and/or swelling of the paws, and a clinical score (maximum of
12/mouse) was assigned as previously described (23). The majority of the GM-CSF1/1 mice exhibited clinical signs of CIA, with
9 of 13 mice having swollen joints beginning on days 24 to 39
(Fig. 1a) and the mean clinical score peaking on day 43 (Fig. 1b).
In Table I, which summarizes the data from Figure 1, it is evident
that the GM-CSF1/1 mice demonstrated a broad range of responses, as we and others have observed using DBA/1 mice in this
model (23, 26); these varied from minor swelling of digits to severe swelling and ankylosis in all four paws (severity range, 1–12;
median, 9).
In contrast, only 2 of 15 GM-CSF2/2 mice responded by day
60, one beginning later than day 45, a significantly lower CIA
incidence than for the GM-CSF1/1 mice ( p , 0.005; see Fig. 1a).
The arthritis in the GM-CSF2/2 mice was restricted to the swelling of single digits on the affected limbs. The mean clinical score
was consistently lower in the GM-CSF2/2 mice than in the GMCSF1/1 mice throughout the course of the experiment (Fig. 1b),
and this was reflected in the significantly lower average clinical
scores of the mice over the period from days 22– 60 ( p , 0.005;
see Table I). The lower clinical scores of the GM-CSF2/2 mice
compared with those of the GM-CSF1/1 mice were thus a combination of reduced incidence of arthritis as well as considerably
lessened severity of disease in responsive mice, shown by the restricted range of 1 to 2 (Table I).
Interestingly, overall the GM-CSF1/2 mice exhibited an intermediate response between the other two genotypes for both incidence and clinical score (see Fig. 1) with 17 of 28 mice developing
signs of arthritis, although the severity range was comparable to
that in the GM-CSF1/1 mice (Table I).
As an additional measure of CIA severity, the weights of the
GM-CSF1/1 and GM-CSF2/2 mice were compared over the period from CII boost (day 21) to sacrifice (day 60). Over this time
frame the GM-CSF1/1 mice failed to gain weight, while the
weights of the GM-CSF2/2 mice increased by an average of 5%
( p , 0.0001, by paired t test), suggesting reduced morbidity and
severity of disease in the latter genotype.
A further experiment was performed using age- and sexmatched GM-CSF1/1 and GM-CSF2/2 mice (n 5 12) of mixed
C57BL/6 and 129/OLA background (single C57BL/6 cross; see
Materials and Methods). Using this cohort, CIA, manifested as a
solitary swollen digit, was observed in only one GM-CSF2/2
mouse. Again, the severity (based on maximum clinical score) of
Histological Grade
3642
PROTECTION FROM CIA IN GM-CSF NULL MICE
FIGURE 3. Humoral response to CII in GM-CSF2/2 and GM-CSF1/1
mice: IgG isotypes. Circulating levels of CII-specific Abs (IgG1, IgG2a,
IgG2b, and IgG3 isotypes) were determined in individual sera from GMCSF2/2 (n 5 15) and GM-CSF1/1 (n 5 13) mice at 60 days postimmunization with CII. Results show the mean 6 SEM (arbitrary units per
milliliter; see Materials and Methods for definition). No significant difference was observed between the two mouse genotypes for levels of any of
the anti-CII IgG isotypes (by Student’s t test).
Cellular response to CII of GM-CSF-deficient mice
Discussion
CIA has been widely used as a means of examining the roles of
cytokines and other inflammatory mediators in arthritis progression. The approaches used have involved direct injection of cytokines, neutralizing Abs, or receptor antagonists of the cytokines.
We previously employed the first of these approaches to demonstrate that GM-CSF exacerbates CIA in DBA/1 mice (23), thereby
illustrating the effects of increased circulating GM-CSF levels on
arthritis development. Numerous studies have employed the second approach (neutralizing Abs) in attempts to demonstrate the
role of targeted endogenous cytokines in CIA, but this approach
FIGURE 2. Histopathology of joints of GM-CSF2/2 and GM-CSF1/1
mice. At 60 days post-primary immunization with CII, mice were sacrificed, their hind limbs were removed, and the paws were processed for
histology as described in Materials and Methods. Frontal sections of the
interphalangeal joints of representative GM-CSF2/2 (a) and GM-CSF1/1
(b and c) mice are shown. The majority of GM-CSF2/2 mouse joints examined were normal in appearance, with smooth intact articular cartilage
(C) and the absence of inflammatory cell infiltrate. The joints of GMCSF1/1 mice most frequently showed severe pathology, with cartilage
erosion down to the subchondral bone (arrowhead), synovial inflammation
(S), and formation of invasive pannus (P) resulting in severe cartilage and
bone degradation (arrows). New bone (NB) formation and remodeling
were also evident in late stage disease. The pannus (see closeup in c)
comprised a mixture of monocyte/macrophages, polymorphonuclear leukocytes, and fibroblast-like cells. J, joint space; B, bone. Hematoxylin and
eosin stain; original magnification, 3100 for a and b and 3250 for c.
level of investigation did not reveal any statistically significant
differences between the two groups of mice in the Ab responses to
CII (Fig. 3).
FIGURE 4. Cellular (DTH) response to CII in GM-CSF2/2 and GMCSF1/1 mice. Age- and sex-matched GM-CSF2/2 and GM-CSF1/1 mice
were immunized by intradermal injection of CII and challenged 10 days
later by s.c. injection into the rear footpad with either CII or vehicle control. Results show the DTH response measured as the difference in millimeters between the increase in thickness of the right (CII-treated) and the
left (vehicle-treated) footpads at 24 and 48 h postchallenge. Values are the
mean 6 SEM for six mice per group. p, p , 0.05 compared with the
GM-CSF1/1 group.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
The cellular response to CII was next investigated by the DTH
reaction based on footpad swelling. While GM-CSF2/2 mice were
capable of a DTH response to CII, as indicated by the significantly
greater degree of swelling in their CII-injected footpads vs that in
vehicle-injected footpads ( p , 0.005, by paired t test; data not
shown), there was a significant reduction ( p , 0.05) in its magnitude compared with that in GM-CSF1/1 mice at 24 and 48 h
after Ag challenge (Fig. 4). Histologic assessment of footpad sections from each group of mice confirmed that the swelling observed was indeed due to mononuclear cell infiltration into the
dermis and not simply edema.
The Journal of Immunology
the two GM-CSF-deficient mice that developed CIA was the first
mouse to show signs of joint swelling (on day 22). Instead, in
accordance with the work of others (21, 33), it would appear that
the capacity of mice to raise Abs to CII is alone insufficient to
result in CIA.
In contrast, while the GM-CSF-deficient mice were capable of
eliciting cell-mediated immunity to CII, as determined by the DTH
response, it was at a significantly reduced level compared with that
in the wild-type control mice (Fig. 4), suggesting suppression of T
cell function. Whether this reduction is sufficient to account for the
minimal arthritic response in these mice remains uncertain. Recent
studies (34) examining the T cell function of immunized GM-CSFdeficient mice reported a reduced CD41 proliferative response to
specific Ags; the mechanism was thought to involve a GM-CSFinduced dendritic cell-derived factor capable of enhancing the T
cell proliferative response, rather than an intrinsic malfunction in
the T cells. The reported incapacity of dendritic cells to either
process or present CII (35) suggests that another explanation may
be needed for the reduced cellular response to CII in the present
study. For instance, it could reflect the inability of the GM-CSFdeficient mouse T cells to activate macrophages through GM-CSF
production. Finally, it must be acknowledged that these studies
compared the immune responses to chick CII; the possibility remains that differences exist between the GM-CSF-deficient and
wild-type mice in their abilities to develop an autoimmune reaction
to murine CII.
The absence of GM-CSF in the knockout mice could, based on
in vitro studies, have ramifications for granulocyte, monocytemacrophage, or dendritic cell responses in the CIA model. Interestingly, GM-CSF-deficient mice have normal hemopoiesis up to
12 wk of age (24) and normal levels of both myeloid-related and
lymphoid-related dendritic cells in the major lymphoid organs
(spleen, lymph node, and thymus) (36). However, these are under
steady state conditions where mice have not been elicited to develop autoimmune inflammatory disease, such as CIA. Local or
systemic differences between the GM-CSF1/1 and GM-CSF2/2
mice in either the numbers or levels of activation of any of the
three cell lineages listed above may still account for the observations reported herein. Recent functional studies with the GM-CSFdeficient mouse have demonstrated an increased tolerance to endotoxin-mediated septic shock that was related to reduced
circulating levels of the cytokines IFN-g, IL-1a, and IL-6 (37).
Moreover, LPS-stimulated peritoneal macrophages from GMCSF2/2 mice produced less IL-1a and nitric oxide than those from
wild-type mice. Thus, while GM-CSF knockout mice have normal
steady state levels of circulating monocytes, they may be impaired
in their ability to respond to certain stimuli. Studies of the local
(joint) and systemic cytokine levels of CII-immunized GM-CSFdeficient mice may provide further insight into how they are protected from CIA development.
The results of this study taken together with our previous report
of exacerbation of the CIA model by exogenous GM-CSF (23)
support the idea of GM-CSF as a proinflammatory mediator and
provide a strong argument for a pivotal role for GM-CSF in CIA
development and inflammatory joint disease.
Acknowledgments
We thank Jennifer Davis for assistance with the maintenance and breeding
of the mice, and Dr. P. Mark Hogarth (Austin Research Institute, Heidelberg, Australia) for helpful discussion.
References
1. Metcalf, D. 1989. The molecular control of cell division, differentiation commitment and maturation in haemopoietic cells. Nature 339:27.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
can have the following limitations: 1) the Ab may not be accessible
to the site of action of the cytokine; and 2) there may be reduced
efficacy of the neutralizing Ab due to an immune response to this
foreign protein in the mice. One way to avoid these potential inadequacies is to employ gene knockout mice, which do not express
the mediator of interest. Recently, several studies have appeared in
the scientific literature using this approach in the CIA model with
various target genes (28 –32), and these have invariably involved
backcrossing the knockout mice onto the CIA-susceptible DBA/1
(H-2q) strain for up to five generations.
In the present study, GM-CSF knockout mice on a C57BL/6
(H-2b) background were employed to test the requirement for endogenous GM-CSF in the development of CIA. We found that
70% of wild-type mice on this background developed CIA within
40 days of primary CII immunization (Fig. 1a, GM-CSF1/1). This
discovery dispensed with the putative requirement to backcross the
GM-CSF null mice onto the DBA/1 background to attain a CIAsusceptible strain. By comparing, in the CIA model, GM-CSFdeficient and wild-type control mice on this background we
showed that the absence of the GM-CSF gene product protected
against disease development. Only 2 of 15 GM-CSF-deficient mice
developed very mild clinical symptoms of CIA, and this was
shown histologically to be confined to the interphalangeal and
metatarsal-phalangeal joints of the digits. In contrast, the GMCSF-competent mice showed a significantly greater incidence of
disease and clinical responses ranging from mild (score 1) to severe (score 12), with swelling and ankylosis of all four paws.
Histologic analyses of arthritic paws confirmed the clinical assessments: the joints of the GM-CSF-deficient mice were most
often normal in histologic appearance, whereas those of the wildtype mice were frequently severely disrupted, with pannus tissue
formation and associated cartilage and bone loss (Table II and Fig.
2). Interestingly, in 2 of the 57 joints examined from the GM-CSFdeficient mice the degree of joint damage was comparable to that
observed in the most severely affected wild-type mice. It would
therefore appear that GM-CSF is not absolutely required for the
development of CIA and that in a small number of individuals
disease within isolated joints can proceed to the end stage in its
absence. Rather, GM-CSF may be needed for the rapid systemic
progression of CIA toward polyarthritis, since during the course of
this study (up to 60 days) arthritis appeared to be confined to
isolated digits on the affected limbs of GM-CSF-deficient mice. In
further delineating the role of GM-CSF in this model it would be
of interest to compare the effects of systemic and local GM-CSF
reconstitution in the GM-CSF null mice.
Investigations were undertaken to examine why GM-CSF-deficient mice were resistant to CIA induction. Given the importance
of GM-CSF in dendritic cell development in vitro (2) and the reported dependence of the CIA model on both humoral and cellmediated immunities to CII (21), GM-CSF-deficient and wild-type
mice were compared for their immune responses to CII. No differences were observed between the two genotypes in the serum
levels of either the total IgG or IgG subclass responses to CII at 60
days postprimary immunization, suggesting a normal humoral response to this Ag in the GM-CSF-deficient mice. Since complement fixing Abs of the IgG2 subclass are considered critical for
CIA development (27), the latter result precluded the possibility
that, although the total IgG levels were comparable between the
two genotypes, other IgG subclasses may have been generated by
the GM-CSF2/2 mice at the expense of the IgG2 subclasses,
thereby accounting for their reduced CIA incidence. The further
possibility that a temporal difference between the humoral responses to CII of the two genotypes may account for the subsequent differences in CIA development is unlikely given that one of
3643
3644
20. Courtenay, J. S., M. J. Dallman, A. D. Dayan, A. Martin, and B. Mosedale. 1980.
Immunisation against heterologous type II collagen induces arthritis in mice.
Nature 283:666.
21. Seki, N., Y. Sudo, T. Yoshioka, S. Sugihara, T. Fujitsu, S. Sakuma, T. Ogawa,
T. Hamaoka, H. Senoh, and H. Fujiwara. 1988. Type II collagen-induced murine
arthritis. I. Induction and perpetuation of arthritis require synergy between humoral and cell-mediated immunity. J. Immunol. 140:1477.
22. Wooley, P. H. 1988. Collagen-induced arthritis in the mouse. Methods Enzymol.
162:361.
23. Campbell, I. K., A. Bendele, D. A. Smith, and J. A. Hamilton. 1997. Granulocytemacrophage colony stimulating factor exacerbates collagen induced arthritis in
mice. Ann. Rheum. Dis. 56:364.
24. Stanley, E., G. J. Lieschke, D. Grail, D. Metcalf, G. Hodgson, J. A. M. Gall,
D. W. Maher, J. Cebon, V. Sinickas, and A. R. Dunn. 1994. Granulocyte/macrophage colony-stimulating factor-deficient mice show no major perturbation of
hematopoiesis but develop a characteristic pulmonary pathology. Proc. Natl.
Acad. Sci. USA 91:5592.
25. Williams, R. O., M. Feldmann, and R. N. Maini. 1992. Anti-tumor necrosis factor
ameliorates joint disease in murine collagen-induced arthritis. Proc. Natl. Acad.
Sci. USA 89:9784.
26. Caccese, R. G., J. L. Zimmerman, and R. P. Carison. 1992. Bacterial lipopolysaccharide potentiates type II collagen-induced arthritis in mice. Med. Inflam.
1:273.
27. Watson, W. C., and A. Townes. 1985. Genetic susceptibility to murine collagen
II autoimmune arthritis: proposed relationship to the IgG2 autoantibody subclass
response, complement C5, major histocompatibility complex (MHC) and nonMHC loci. J. Exp. Med. 162:1878.
28. Vermeire, K., H. Heremans, M. Vandeputte, S. Huang, A. Billiau, and
P. Matthys. 1997. Accelerated collagen-induced arthritis in IFN-g receptor-deficient mice. J. Immunol. 158:5507.
29. Mori, L., S. Iselin, G. D. Libero, and W. Lesslauer. 1996. Attenuation of collagen-induced arthritis in 55-kDa TNF receptor type 1 (TNFR1)-IgG1-treated and
TNFR1-deficient mice. J. Immunol. 157:3178.
30. Manoury-Schwartz, B., G. Chiocchia, N. Bessis, O. Abehsira-Amar, F. Batteux,
S. Muller, S. Huang, M.-C. Boissier, and C. Fournier. 1997. High susceptibility
to collagen-induced arthritis in mice lacking IFN-g receptors. J. Immunol. 158:
5501.
31. McIntyre, K. W., D. J. Shuster, K. M. Gillooly, R. R. Warrier, S. E. Connaughton,
L. B. Hall, L. H. Arp, M. K. Gately, and J. Magram. 1996. Reduced incidence and
severity of collagen-induced arthritis in interleukin-12-deficient mice. Eur. J. Immunol. 26:2933.
32. Bullard, D. C., L. A. Hurley, I. Lorenzo, L. M. Sly, A. L. Beaudet, and
N. D. Staite. 1996. Reduced susceptibility to collagen-induced arthritis in mice
deficient in intercellular adhesion molecule-1. J. Immunol. 157:3153.
33. Verdrengh, M., R. Holmdahl, and A. Tarkowski. 1995. Administration of antibodies to hyaluronan receptor (CD44) delays the start and ameliorates the severity of collagen II arthritis. Scand. J. Immunol. 42:353.
34. Wada, H., Y. Noguchi, M. W. Marino, A. R. Dunn, and L. J. Old. 1997. T cell
functions in granulocyte/macrophage colony-stimulating factor deficient mice.
Proc. Natl. Acad. Sci. USA 94:12557.
35. Michaëlsson, E., M. Holmdahl, Å. Engström, H. Burkhardt, A. Scheynius, and
R. Holmdahl. 1995. Macrophages, but not dendritic cells, present collagen to T
cells. Eur. J. Immunol. 25:2234.
36. Vremec, D., G. J. Lieschke, A. R. Dunn, L. Robb, D. Metcalf, and K. Shortman.
1997. The influence of granulocyte/macrophage colony-stimulating factor on
dendritic cell levels in mouse lymphoid organs. Eur. J. Immunol. 27:40.
37. Basu, S., A. R. Dunn, M. W. Marino, H. Savoia, G. Hodgson, G. J. Lieschke, and
J. Cebon. 1997. Increased tolerance to endotoxin by granulocyte-macrophage
colony-stimulating factor-deficient mice. J. Immunol. 159:1412.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
2. Inaba, K., M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu,
and R. M. Steinman. 1992. Generation of large numbers of dendritic cells from
mouse bone marrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 176:1693.
3. Hamilton, J., E. Stanley, A. Burgess, and R. Shadduck. 1980. Stimulation of
macrophage plasminogen activator activity by colony-stimulating factors. J. Cell.
Physiol. 103:435.
4. Gamble, J., M. Elliot, E. Jaipargas, A. Lopez, and M. Vadas. 1989. Regulation of
human monocyte adherence by granulocyte-macrophage colony-stimulating factor. Proc. Natl. Acad. Sci. USA 86:7169.
5. Alvaro-Gracia, J. M., N. J. Zvaifler, and G. S. Firestein. 1989. Cytokines in
chronic inflammatory arthritis. IV. Granulocyte/macrophage colony-stimulating
factor-mediated induction of class II MHC antigen on human monocytes: a possible role in rheumatoid arthritis. J. Exp. Med. 170:865.
6. Arnaout, A., E. Wang, S. Clark, and C. Sief. 1986. Human recombinant granulocyte-macrophage colony-stimulating factor increases cell-to-cell adhesion and
surface expression of adhesion-promoting surface glycoproteins on mature granulocytes. J. Clin. Invest. 78:597.
7. Morrissey, P. J., L. Bressler, L. S. Park, A. Alpert, and S. Gillis. 1987. Granulocyte-macrophage colony-stimulating factor augments the primary antibody response by enhancing the function of antigen-presenting cells. J. Immunol. 139:
1113.
8. Hart, P. H., G. A. Whitty, D. S. Piccoli, and J. A. Hamilton. 1988. Synergistic
activation of human monocytes by granulocyte-macrophage colony-stimulating
factor and IFN-g: increased TNF-a but not IL-1 activity. J. Immunol. 141:1516.
9. Sisson, S. D., and C. A. Dinarello. 1988. Production of interleukin-1a, interleukin-1b and tumor necrosis factor by human mononuclear cells stimulated with
granulocyte-macrophage colony-stimulating factor. Blood 72:1368.
10. Fleischmann, J., D. Golde, R. Weisbart, and J. Gasson. 1986. Granulocyte-macrophage colony-stimulating factor enhances phagocytosis of bacteria by human
neutrophils. Blood 68:708.
11. Weisbart, R. H., L. Kwan, D. W. Golde, and J. C. Gasson. 1987. Human GMCSF primes neutrophils for enhanced oxidative metabolism in response to the
major physiological chemoattractants. Blood 69:18.
12. Gasson, J., R. Weisbart, S. Kauffman, S. Clarke, R. Hewick, and G. Wong. 1984.
Purified human granulocyte-macrophage colony-stimulating factor: direct action
on neutrophils. Science 226:1339.
13. Xu, W. d., G. S. Firestein, R. Taetle, K. Kaushansky, and N. J. Zvaifler. 1989.
Cytokines in chronic inflammatory arthritis. II. Granulocyte-macrophage colonystimulating factor in rheumatoid synovial effusions. J. Clin. Invest. 83:876.
14. Campbell, I. K., U. Novak, J. Cebon, J. E. Layton, and J. A. Hamilton. 1991.
Human articular cartilage and chondrocytes produce hemopoietic colony-stimulating factors in culture in response to IL-1. J. Immunol. 147:1238.
15. Leizer, T., J. Cebon, J. E. Layton, and J. A. Hamilton. 1990. Cytokine regulation
of colony-stimulating factor production in cultured human synovial fibroblasts. I.
Induction of GM-CSF and G-CSF production by interleukin-1 and tumor necrosis
factor. Blood 76:1989.
16. Hamilton, J. A. 1993. Rheumatoid arthritis: opposing actions of haemopoietic
growth factors and slow-acting anti-rheumatic drugs. Lancet 342:536.
17. Kowanko, I. C., and A. Ferrante. 1991. Granulocyte-macrophage colony-stimulating factor augments neutrophil-mediated cartilage degradation and neutrophil
adherence. Arthritis Rheum. 34:1452.
18. de Vries, E., P. Willemse, B. Biesma, A. Stern, P. Limburg, and E. Vellenga.
1991. Flare-up of rheumatoid arthritis during GM-CSF treatment after chemotherapy. Lancet 338:517.
19. Hazenberg, B., M. V. Leeuwen, M. V. Rijswijk, A. Stern, and E. Vellenga. 1898.
Correction of granulocytopenia in Felty’s syndrome by granulocyte-macrophage
colony-stimulating factor: simultaneous induction of interleukin-6 release and
flare-up of the arthritis. Blood 74:2769.
PROTECTION FROM CIA IN GM-CSF NULL MICE